Expert System for Determining the Level of Stress before Pediatric Dental Treatment Procedia Technology 12 ( 2014 ) 548 – 557 2212-0173 © 2013 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Selection and peer-review under responsibility of the Petru Maior University of Tirgu Mures. doi: 10.1016/j.protcy.2013.12.528 The 7th International Conference Interdisciplinarity in Engineering (INTER-ENG 2013) Expert system for determining the level of stress before pediatric dental treatment Sorana-Maria Bucura,*, Manuela Chibeleana, Adrian Gligorb, Mariana Pacurara a University of Medicine and Pharmacy Tg-Mures, 38 Gh. Marinescu Street, Tg-Mures 540139, Romania b“Petru Maior” University, 1 Nicolae Iorga street, 540088 Tirgu Mures, Romania Abstract The paper is focused on finding a solution based on an expert system designed to ease the dentist’s treatment decision by evaluating the patient’s anxiety and stress status. To identify the requirements and functionalities of this system there was performed an extensive research based on psychological questionnaires and the use of a powerful device made to determine body energy and stress variables. The experimental results were summarized and conclusions related to the proposed solution were discussed. Then we presented the software specially designed to establish the opportunity of clinical intervention considering the patient’s stress and anxiety. © 2013 The Authors. Published by Elsevier B.V. Selection and peer-review under responsibility of Department of Electrical and Computer Engineering, Faculty of Engineering, “Petru Maior” University of Tîrgu Mureș. Keywords: expert system; medical decision support system for dentists; anxiety; dental treatment. 1. Introduction Before initiating any dental treatment it is relevant to evaluate the clinical cooperation of the patients [1]. Getting a successful result in pediatric dentistry needs a good communication with the young patient. This is not possible when fear intervenes in the confrontation with the doctor [2]. * Corresponding author. Tel.: +0-040-753-756-551 E-mail address: bucur_sorana@yahoo.com Available online at www.sciencedirect.com © 2013 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Selection and peer-review under responsibility of the Petru Maior University of Tirgu Mures. ScienceDirect http://creativecommons.org/licenses/by-nc-nd/3.0/ http://creativecommons.org/licenses/by-nc-nd/3.0/ 549 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 Less compliance could be easily associated to the person’s fear. Any dentist could expect a certain anxiety or fear from the young patient but these become a problem when they are exaggerated [1, 2]. Many patients faced with unpleasant or painful past experiences consequently manifest an unjustified fear when confronted with the doctor, the dental equipment, different smells of the dental office or sounds produced by dental drills. Exaggerated fear can lead to poor collaboration with the dentist, a difficult treatment, bad results and undesirable events [3, 4]. The goal of the study was to determine the exact anxiety level in patients using a computer program specially designed for this purpose. This software establishes the opportunity of medical intervention considering the patient’s anxiety and related symptoms. 2. Theoretical background Anxiety includes state anxiety which is defined as “a transitory emotional condition that varies in intensity and fluctuates over time” whereas” trait anxiety is a personality trait that remains relatively stable“[5]. It refers to the tendency to be anxious without external stress. First, we determined the patient’s anxiety level by using a psychological questionnaire, Spielberger’s State-Trait Anxiety Inventory - STAI - which has been approved and validated in many studies [6]. We have adapted this questionnaire in order to better match our study. The Romanian translation of the questionnaire was used. STAI is composed of two scales, A-state and A-trait, measuring distinct anxiety concepts: state anxiety (how someone is feeling in a certain condition) and trait anxiety (common fear of someone). ” Both scales consist of twenty items for which a person rates anxiety on a scale from one (almost never) to four (very much so)” [5]. It has been shown that A-state scores increase after various kinds of stress and decrease after relaxation training [7]. Therefore, this scale can be used to measure changes in the intensity of the anxious condition occurring in certain situations. Generally, those who score high in A-trait will show increases in A-state more frequently than individuals who score low in A-trait because they tend to react to a large number of cases, judging them dangerous or threatening [8]. State anxiety is a transitory emotional state or condition of the body characterized by subjective feelings, consciously perceived tension and fear and increased activity of ANS-Autonomic Nervous System [9]. This nervous system controls vital functions such as heart activity, blood pressure, digestion, salivation, perspiration and exchanges between body and environment [10]. The sympathetic nervous system plays a role in stress situations with all aspects of increased secretion of adrenaline: peripheral vasodilatation, acceleration of heart rate, rise in blood pressure, sweating, hyper salivation; skeletal muscles are well supplied with blood at the expense of internal organs, the pupil of the eye is increased, the whole body being in a state of alarm. The second step of the study was to determine the stress status of the patient with a powerful device used in traditional Chinese medicine and homeopathy [11] approved in E.U. Over the last decades scientists have shown that electromagnetic fields govern biological systems. The device measures the energy level of the whole body shown by Energy parameter which is the average of energy values determined in 24 energy points located on the skin surface of hands and feet, corresponding to various organs and systems. It also evaluates the stress status of the autonomic nervous system associated with symptoms and gives us indications of potential risks for the body health when these values are increased [12]. 3. Research methodology In order to investigate the correspondence between the values given by anxiety questionnaires and those shown by device’s parameters we formulated the following hypotheses: Research hypotheses: We assume that changes in the normal values of trait anxiety can lower the body energy shown by Energy parameter of the device and lead to diseases; 550 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 We suppose that high values of state anxiety are related with high values of the ANS parameter which is shown in a strong relation with stress symptoms by the device we used. 3.1. Research variables In the present research we established an independent variable and dependent variables measured with digital and nominal scales, thus: Age of the patients was the independent variable Dependent variables were: State anxiety expressed by A-state score Trait anxiety expressed by A-trait score Energy parameter which shows the energetic level of the body; we called it E for ease of expression. ANS parameter which reveals the stress at a certain moment; we called it ANS for ease of expression because it’s related to the autonomic nervous system activity. 3.2. Methods To undergo this study, we used the following methods: 1. Psychometric methods STAI questionnaire where we adapted A-scale to our research by adding the formulation “because of the state of my teeth” to the items 2, 3, 6, 7, 9, 10, 12, 14, 16-20. It is allowed even indicated by the authors to adapt their A- state scale to a specific situation for more accurate results. energetic measurements using the device described above 2. Statistical methods descriptive statistical indices correlations The interpretation of the results was performed by the SPSS statistical program. 3.3. Sample of subjects. The research was based on a group of subjects consisting of 30 patients, girls and boys aged between 11 and 18, randomly selected. Toolbox used: clinical observation psychological questionnaire measuring device for highlighting energy level and stress status 3.4. Working procedure The experimental study began with a clinical examination of each patient. Following discussion and selection, the subjects were classified according to their age and sex. 3.5. Presentation and data analysis Fig.1 shows the patients’ distribution according to sex criterion. 551 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 Fig. 1. Patient’s distribution according to sex criterion chart So the randomly chosen sample contains mostly girls with a difference of 46.67% between girls and boys. Fig.2 shows the distribution of patients according to age. 16,66% 23,33% 23,33% 36,66% 0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 11-12 13-14 15-16 17-18 age % Fig. 2. Distribution of patients according to age From the chart above we can see that the predominant adolescent age group (17-18 year old) represents 36.66% of the subjects investigated. The patients from our sample filled in the Spielberger’s State -Trait Anxiety Inventory STAI with those two scales described above - A-trait and A-state - and the results are presented by the next charts: Fig. 3. Distribution of patients according to the values obtained at A-trait questionnaire From the chart above we see that patients who scored values between 35 and 45 represent 69.99% of the research sample. The patients’ distribution according to the values obtained at A-state questionnaire is presented in Fig.4. Girls 73.33% Boys 26.66% 6.66% 36.66% 33.30% 10% 13.33% Score 30-35 Score 35-40 Score 40-45 Score 45-50 Score 50-75 552 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 6.66% 36.66% 33.30% 10% 13.33% Score 30-35 Score 35-40 Score 40-45 Score 45-50 Score 50-75 Fig. 4. Distribution of patients according to the values obtained at A-state questionnaire Fig.5 shows the distribution of patients according to the energetic level of the body, “E”. Fig. 5. Distribution of patients according to the energetic level of the body, “E” From the above chart we can see that the majority of patients, 59.98% which represents more than a half of the research group, were located between 25-55 normal values of this parameter. The ANS normal values are from 0 to 2. More than 2 means we are dealing with a stress status. The patients’ distribution according to the ANS parameter’s values is presented in Fig.6. 23.33% 26.66% 33.33% 3.33% 3.33% value 1-2 value 2-3 value 3-4 value 4-5 value 6-7 Fig. 6. Distribution of patients according to the ANS parameter 6.66% 26.66% 13.33% 19.99% 20% 6.66% 6.66% value 10-20 value 20-30 value 30-40 value 40-50 value 50-60 value 60-70 553 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 4. Results. Main findings 4.1. Results related to Hypothesis 1 We assume that changes in the normal values of trait anxiety can lower the body energy shown by Energy parameter (E) of the device and lead to diseases. In order to verify this hypothesis we used correlations and t-test from statistical program SPSS. We correlated Variable 2-A trait with Variable 3-E and we saw that their means are relatively close which leads to the conclusion that there is a strong relation between the trait anxiety and the energetic level of the body. Table 1: Means Cases Percent Excluded VAR00002 VAR00001 VAR00003 VAR00001 Included Percent Percent N N 30 49.2% 31 50.8% 100.0% 30 49.2% 31 50.8% 100.0% Legend: Var0001=age, Var0002=-A-trait, Var0003= E Table 2: Relations between Variable 2 and Variable 3 Total Mean 42.7333 41.4667 N 30 30 Std. Deviation 6`.4108 16.4981 Correlations Table 3: Correlations between Variable 2=A-trait and Variable 3=A-state VAR00002 VAR00003 VAR00002 Pearson Correlation 1.000 .342 Sig. (2-tailed) . .064 N 30 30 VAR00003 Pearson Correlation .342 1.000 Sig. (2-tailed) .064 . N 30 30 Legend : Pearson=correlation, Var 0002=Variable A-trait, Var0003=Variable E One-Sample Statistics Table 4: t-test N Mean Std. Deviation Std. Error Mean VAR00002 30 42.7333 6.4108 1.1705 VAR00003 30 41.4667 16.4981 3.0121 Legend : Variable 1=age, Variable 2=A-trait, Variable3=E, N=subjects t= -0.342 statistically significant at a significance of 0.05, which shows that the difference is significant and hypothesis 1 is valid. So, we found out by statistical methods that hypothesis 1 is valid. 4.2. Results related to Hypothesis 2 We suppose that high values of state anxiety are related with high values of the ANS parameter which is shown by the device used in a strong relation with stress symptoms. 554 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 Case processing summary: Table 5: Means Cases Included Excluded Total N Percent N Percent N Percent VAR00001 VAR00002 30 100.0% 0 .0% 30 100.0% We correlated Variable 2-A trait with Variable 3-E and we saw that their means are relatively close which leads to the conclusion that there is a strong relation between the trait anxiety and the energetic level of the body. Table 6: Report VAR00001 VAR00002 Mean N Std. Deviation Total 48.2000 30 8.2813 From the table above we see a standard deviation of 8.2813 between the two variables A-state and ANS their mean being 48.200 One-Sample Statistics Table 7: t-test N Mean Std. Deviation Std. Error Mean VAR00001 30 48.2000 8.2813 1.5119 VAR00002 30 2.7333 1.0694 .1952 We see that mean between the two variables is 48.200 bigger than 45.4667 and standard deviation is 1.3167, which means that there is a connection between the stress status shown by ANS parameter of the device and anxiety as a state evidenced by A-state questionnaire. Table 8: One-Sample Test Test Value = 0 t df Sig. (2-tailed) Mean Difference 95% Confidence Interval of the Difference Lower Upper VAR00001 31.879 29 .000 48.2000 45.1077 51.2923 VAR00002 14.000 29 .000 2.7333 2.3340 3.1326 Table 9: Correlations VAR00001 VAR00002 VAR00001 Pearson Correlation 1.000 .667 Sig. (2-tailed) . .000 N 30 30 VAR00002 Pearson Correlation .667 1.000 Sig. (2-tailed) .000 . N 30 30 ** Correlation is significant at the 0.01 level (2-tailed). Legend : Variable 0001=A-state, Var0002=ANS 555 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 t = -0.667 statistically significant at a significance less than 0.01 which shows that the difference is significant and hypothesis1 is valid. So, we found out by statistical methods that hypothesis 2 is valid. 5. System implementation and testing We choose to implement a web based expert system prototype. The architecture of the application is shown in Fig. 7. Fig. 7. Application architecture. We determined strong correlations between A-trait and E, A-state and ANS; those allowed the design of an expert system developed to ease treatment decision. Following we present the knowledge base of the expert system built to assist the dentist to take the most adequate therapeutic decision [13, 14, 15]. The rule base of the developed expert system is built on the following [16]: If E is between 25-55 and ANS less than 2, we start the treatment because stress status of the patient is normal. If E is less than 25 and ANS less than 2, we start the treatment because state anxiety has normal values but we suspect chronic diseases due to the lack of the body energy; so we advice the patient to see a doctor. If E is more than 55 and ANS less than 2, we start our treatment very carefully because any time the patient may become stressed. If E is between 25-55 and ANS more than 2, we postpone our treatment and apply psychological relaxation techniques. If E is less than 25 and ANS more than 2, we postpone our treatment and apply psychological relaxation techniques. We advice the patient to investigate his/her health because of lack of the body energy. If E is more than 55 and ANS more than 2, the treatment is compromised because of the exaggerated stress status of the patient. The system was tested extensively, using historical medical records. The small number of rules and their semantic consistency generates fast and reliable solutions, with real support for the decision-making process. For example: Patient A.I., female, 13 years, E=35, ANS=2.74, we postponed our treatment and applied psychological relaxation techniques. Then we could successfully complete treatment. Patient S.V., male, 14 years, E=22, ANS=1.71, we applied the treatment without any problems but in anamnesis the child reported diabetes. Application Server Inference module Storage Web-based client Security and access module Domain & Knowledge Database 556 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 Patient B.I., female, 17 years, E=70, ANS=3.56, we could not even start our treatment because of the patient’s high level of stress. Patient T.R., male, 15 years, E=67, ANS=1.87, we finished our treatment with some difficulties. Patient A.S., female, 17 years, E=45, ANS=1.62, we completed our treatment without any problems. Patient T.R., male, 16 years, E=21, ANS=3.11, we postponed our treatment and applied psychological relaxation techniques. We advised the patient to investigate his health because of lack of the body energy. 6. Conclusions The present work was carried out a research on anxiety which led to the development of a medical decision support application. We found out that changes in the normal values of trait anxiety can lower the body energy shown by Energy parameter of the device and lead to diseases. We found that high values of state anxiety are related with high values of the ANS parameter which is shown in a strong relation with stress symptoms by the device used. So, we can correlate the two device’s parameters values with those given by the most common and used anxiety questionnaire, practically we validated the device’s measurements. Then became possible to develop a software application designed to ease doctor’s work by helping him with the treatment decision. This expert system greatly eases clinical work by helping the dentist to take the best medical decision at the beginning of the young patient’s treatment. Acknowledgement This study was entirely supported by the Private Medical Office belonging to Dr. Bucur Sorana-Maria and approved by Decision 95 of the Ethics Committee of the University of Medicine and Pharmacy Tg-Mures. We respected the confidentiality of personal data of the patients. The parents of the minor patients were informed about the study and agreed with their children participation by writing an informed consent. They agreed with the use of the data obtained in scientific purpose and for publication. We are grateful to the office staff for their help in conducting the study. The paper was partly supported by the Sectorial Operational Programme Human Resources Development (SOP HRD), financed from the European Social Fund and by the Romanian Government under the contract number POSDRU 80641. References [1] Vassend O, Willumsen T, Hoffart A. Effects of Dental Fear Treatment on General Distress: The Role of Personality Variables and Treatment Method. Behav Modif 2000, 24: 580-599. [2] Townend E, Dimigen G, Fung D. A clinical study of child dental anxiety. Behav Res Ther 2000, 38: 31-46. [3] Kent G, Warren P. A Study of Factors Associated with Changes in Dental Anxiety. J of Dental Research, November 1, 1985, 64: 1316-1318. [4] Arnrup K, Broberg AG, Berggren U, Bodin L. Temperamental reactivity and negative emotionality in uncooperative children referred to specialized paediatric dentistry compared to children in ordinary dental care. Int J Paediatric Dentistry 2007, 17: 419-429. [5] Spielberger CD, Gorsuch RL, Lushene R. Manual for the State-Trait Anxiety Inventory. Consulting Psychologists Press, Palo Alto, California, 1983. [6] Okun A, Stein RE, Bauman LJ, Silver EJ. Content validity of the State-Trait Anxiety Inventory from the perspective of DSM-IV. Psychol Rep 2006, 79: 1060-1069. [7] Eppley KR, Abrams AI, Shear J. Differential effects of relaxation techniques on trait-anxiety: A meta-analysis. J of Clinical Psychology, 1989, 45: 957-974 . [8] Grossman P, Niemann L, Schmidt S, Walach H. Mindfulness – based stress reduction and stress benefits: A meta-analysis. J of Psychosomatic Research 2004, 57: 35-43. [9] West GA, Reid KH, Bastawi AE. Clinical Science: Autonomic Responses to Dental Procedures in Pedodontic Patients During a Standard Restoration Session. J of Dental Research 1985 62: 728-732. [10] Watson D, Pennebaker, JW.Health complaints, stress and distress: Exploring the central role of negative affectivity. Psychol Review, 1989, 96(2): 234-255. 557 Sorana-Maria Bucur et al. / Procedia Technology 12 ( 2014 ) 548 – 557 [11] Sherman KI, Hogeboom CI, Cherkin DC. How traditional Chinese medicine acupuncturists would diagnose and treat chronic low back pain: results of a survey of licensed acupuncturists in Washington State. Complementary Therapies in Medicine, 2001, 9: 146-153. [12] Boss A, Hoogstraten J, Prahl-Andersen B. On the use of personality characteristics in predicting compliance in orthodontic practice. Am J Orthod Dentofacial Orthop 2003, 123: 569-572. [13] Russello G, Chaudron M R, van Steen M, Bokharouss I – An experimental evaluation of self-managing availability in shared data spaces. Science of Computer Programming 2007, 64: 246-262. [14] Krikhaar R, Crukovic I – Software Configuration Management. Science of Computer Programming 2007, 65: 215-221. [15] Colman A, Han J - Using role-based coordination to achieve software adaptability. Science of Computer Programming 2007, 64: 223-245. [16] American Academy of Pediatric Dentistry - Guidelines for behavior management. Pediatr Dent, Supplement issue 2002-2003, 24: 68- 74.14Krikhaar R, Crukovic I – Software Configuration Management. Science of Computer Programming 2007, 65: 215-221.